anysite-cli 0.1.5__py3-none-any.whl → 0.1.7__py3-none-any.whl

This diff represents the content of publicly available package versions that have been released to one of the supported registries. The information contained in this diff is provided for informational purposes only and reflects changes between package versions as they appear in their respective public registries.
@@ -315,8 +315,14 @@ class DatasetDbLoader:
315
315
  return _extract_dot_value(record, diff_key)
316
316
  return record.get(diff_key)
317
317
 
318
+ # Build field mapping for db_load.fields filtering
319
+ field_mapping = self._get_db_field_mapping(source)
320
+
318
321
  # Determine the DB column name for the key
319
- db_key_col = diff_key.replace(".", "_")
322
+ if field_mapping and diff_key in field_mapping:
323
+ db_key_col = field_mapping[diff_key]
324
+ else:
325
+ db_key_col = diff_key.replace(".", "_")
320
326
 
321
327
  # INSERT added records
322
328
  if result.added:
@@ -326,13 +332,14 @@ class DatasetDbLoader:
326
332
  total += 1
327
333
 
328
334
  # DELETE removed records (skipped in append mode)
335
+ ph = self._placeholder()
329
336
  if result.removed and sync_mode == "full":
330
337
  safe_col = sanitize_identifier(db_key_col)
331
338
  for record in result.removed:
332
339
  key_val = _get_key_val(record)
333
340
  if key_val is not None:
334
341
  self.adapter.execute(
335
- f"DELETE FROM {table_name} WHERE {safe_col} = ?",
342
+ f"DELETE FROM {table_name} WHERE {safe_col} = {ph}",
336
343
  (str(key_val),),
337
344
  )
338
345
  total += 1
@@ -348,20 +355,34 @@ class DatasetDbLoader:
348
355
  if not changed_fields:
349
356
  continue
350
357
 
351
- # Build SET clause from changed fields
358
+ # Build SET clause only include fields that exist in the DB
352
359
  set_parts = []
353
360
  params: list[Any] = []
354
361
  for field_name in changed_fields:
355
- new_val = record.get(field_name)
356
- safe_field = sanitize_identifier(field_name)
357
- set_parts.append(f"{safe_field} = ?")
362
+ if field_mapping is not None:
363
+ if field_name not in field_mapping:
364
+ continue
365
+ db_col = field_mapping[field_name]
366
+ else:
367
+ db_col = field_name
368
+
369
+ if "." in field_name:
370
+ new_val = _extract_dot_value(record, field_name)
371
+ else:
372
+ new_val = record.get(field_name)
373
+
374
+ safe_field = sanitize_identifier(db_col)
375
+ set_parts.append(f"{safe_field} = {ph}")
358
376
  params.append(new_val)
359
377
 
378
+ if not set_parts:
379
+ continue
380
+
360
381
  params.append(str(key_val))
361
382
  sql = (
362
383
  f"UPDATE {table_name} "
363
384
  f"SET {', '.join(set_parts)} "
364
- f"WHERE {safe_col} = ?"
385
+ f"WHERE {safe_col} = {ph}"
365
386
  )
366
387
  self.adapter.execute(sql, tuple(params))
367
388
  total += 1
@@ -375,6 +396,36 @@ class DatasetDbLoader:
375
396
  return other.dependency.field
376
397
  return None
377
398
 
399
+ def _get_db_field_mapping(self, source: DatasetSource) -> dict[str, str] | None:
400
+ """Build mapping of parquet_field -> db_column from db_load.fields.
401
+
402
+ Returns None if no explicit fields configured (all fields allowed).
403
+ """
404
+ db_load = source.db_load
405
+ if not db_load or not db_load.fields:
406
+ return None
407
+
408
+ mapping: dict[str, str] = {}
409
+ for field_spec in db_load.fields:
410
+ alias = None
411
+ upper = field_spec.upper()
412
+ as_idx = upper.find(" AS ")
413
+ if as_idx != -1:
414
+ alias = field_spec[as_idx + 4:].strip()
415
+ source_field = field_spec[:as_idx].strip()
416
+ else:
417
+ source_field = field_spec
418
+
419
+ col_name = alias or source_field.replace(".", "_")
420
+ mapping[source_field] = col_name
421
+ return mapping
422
+
423
+ def _placeholder(self) -> str:
424
+ """Get the parameter placeholder for the dialect."""
425
+ if self._dialect == "postgres":
426
+ return "%s"
427
+ return "?"
428
+
378
429
  def _auto_id_type(self) -> str:
379
430
  """Get the auto-increment ID column type for the dialect."""
380
431
  if self._dialect == "postgres":
@@ -1,6 +1,6 @@
1
1
  Metadata-Version: 2.4
2
2
  Name: anysite-cli
3
- Version: 0.1.5
3
+ Version: 0.1.7
4
4
  Summary: CLI for Anysite API - web data extraction for humans and AI agents
5
5
  Project-URL: Homepage, https://anysite.io
6
6
  Project-URL: Documentation, https://docs.anysite.io/cli
@@ -21,7 +21,7 @@ anysite/dataset/__init__.py,sha256=J0sKQkGwVOPtvp6pka7LcdeUEADvjWRs71yRuROzJxI,8
21
21
  anysite/dataset/analyzer.py,sha256=8dsPW32SbSaUTy1F0NIed1U45wjiMgQeJ2iWX7hBxRQ,9245
22
22
  anysite/dataset/cli.py,sha256=rEWK1ka-YQ_Vbbj2nMaMYTD9g3wa3ethUWSoaWRSGTY,23066
23
23
  anysite/dataset/collector.py,sha256=ZdR3CmQQew_iuJpNtJ4knSrjt0hvkEL4WIaS0IKEkwQ,23927
24
- anysite/dataset/db_loader.py,sha256=ksvRt-VJISL4Syk2O1-TTkOMj1uGzk7aQARYS2n--U4,13751
24
+ anysite/dataset/db_loader.py,sha256=ebx-SJa-w4EjUsuoNqU2wtAWlbXC93kCc9shs3ZHm70,15597
25
25
  anysite/dataset/differ.py,sha256=jB_VWTb7UuEBWG9nv1ry5xeo9hmWdhA_cTm6Ed43_Uw,17746
26
26
  anysite/dataset/errors.py,sha256=r8cZXoIzSeTGCWpeYjntnN0AduCu74YZyWs3sFu17J4,914
27
27
  anysite/dataset/exporters.py,sha256=mA2FYbYJbHfrwkXbHDu4g5qPG_JJKnkVciXFKPkF1Vw,3708
@@ -58,8 +58,8 @@ anysite/streaming/writer.py,sha256=HfMsC4umUdJuNIAPK57YAxEGyTwUmy-zNrqFkwY6aew,4
58
58
  anysite/utils/__init__.py,sha256=7SnbxpxKENK-2ecUL5NfnZ9okGI7COKYw4WF46172HM,23
59
59
  anysite/utils/fields.py,sha256=bSrHadzNmabL4qubqhXXZoWb_P8KA-3S7_FLVT8nGBc,7410
60
60
  anysite/utils/retry.py,sha256=89TbXvavi5t22P2mTYCLAS6SSZoW65gQ0nnYNbYAF0M,2684
61
- anysite_cli-0.1.5.dist-info/METADATA,sha256=B4HxyrTZxBbhMb17lb0LoRcne_cRehz8xNUYIvDraMA,12437
62
- anysite_cli-0.1.5.dist-info/WHEEL,sha256=WLgqFyCfm_KASv4WHyYy0P3pM_m7J5L9k2skdKLirC8,87
63
- anysite_cli-0.1.5.dist-info/entry_points.txt,sha256=FDPxNasy0fRRcOgJdZRVP7Qw01C3TwRa1OwPJiskNyg,45
64
- anysite_cli-0.1.5.dist-info/licenses/LICENSE,sha256=gVAxkI23CFm4x4HV_fkQYw_bGq93mQmVZEwxNs-YTa4,1069
65
- anysite_cli-0.1.5.dist-info/RECORD,,
61
+ anysite_cli-0.1.7.dist-info/METADATA,sha256=RLzLHvkzQ52E6Ovse5FrXEYjJ2P_6X_sGrs_F5ho2Uk,12437
62
+ anysite_cli-0.1.7.dist-info/WHEEL,sha256=WLgqFyCfm_KASv4WHyYy0P3pM_m7J5L9k2skdKLirC8,87
63
+ anysite_cli-0.1.7.dist-info/entry_points.txt,sha256=FDPxNasy0fRRcOgJdZRVP7Qw01C3TwRa1OwPJiskNyg,45
64
+ anysite_cli-0.1.7.dist-info/licenses/LICENSE,sha256=gVAxkI23CFm4x4HV_fkQYw_bGq93mQmVZEwxNs-YTa4,1069
65
+ anysite_cli-0.1.7.dist-info/RECORD,,